|
1
|
|
|
"""Definitions for REST API operations. These are wired via Connexion to |
|
2
|
|
|
methods defined in the Swagger specification.""" |
|
3
|
|
|
|
|
4
|
|
|
import connexion |
|
5
|
|
|
from simplemma.langdetect import lang_detector |
|
6
|
|
|
|
|
7
|
|
|
import annif.registry |
|
8
|
|
|
from annif.corpus import Document, DocumentList, SubjectSet |
|
9
|
|
|
from annif.exception import AnnifException |
|
10
|
|
|
from annif.project import Access |
|
11
|
|
|
from annif.suggestion import SuggestionFilter |
|
12
|
|
|
|
|
13
|
|
|
|
|
14
|
|
|
def project_not_found_error(project_id): |
|
15
|
|
|
"""return a Connexion error object when a project is not found""" |
|
16
|
|
|
|
|
17
|
|
|
return connexion.problem( |
|
18
|
|
|
status=404, |
|
19
|
|
|
title="Project not found", |
|
20
|
|
|
detail="Project '{}' not found".format(project_id), |
|
21
|
|
|
) |
|
22
|
|
|
|
|
23
|
|
|
|
|
24
|
|
|
def server_error(err): |
|
25
|
|
|
"""return a Connexion error object when there is a server error (project |
|
26
|
|
|
or backend problem)""" |
|
27
|
|
|
|
|
28
|
|
|
return connexion.problem( |
|
29
|
|
|
status=503, title="Service unavailable", detail=err.format_message() |
|
30
|
|
|
) |
|
31
|
|
|
|
|
32
|
|
|
|
|
33
|
|
|
def list_projects(): |
|
34
|
|
|
"""return a dict with projects formatted according to Swagger spec""" |
|
35
|
|
|
|
|
36
|
|
|
return { |
|
37
|
|
|
"projects": [ |
|
38
|
|
|
proj.dump() |
|
39
|
|
|
for proj in annif.registry.get_projects(min_access=Access.public).values() |
|
40
|
|
|
] |
|
41
|
|
|
} |
|
42
|
|
|
|
|
43
|
|
|
|
|
44
|
|
|
def show_project(project_id): |
|
45
|
|
|
"""return a single project formatted according to Swagger spec""" |
|
46
|
|
|
|
|
47
|
|
|
try: |
|
48
|
|
|
project = annif.registry.get_project(project_id, min_access=Access.hidden) |
|
49
|
|
|
except ValueError: |
|
50
|
|
|
return project_not_found_error(project_id) |
|
51
|
|
|
return project.dump() |
|
52
|
|
|
|
|
53
|
|
|
|
|
54
|
|
|
def detect_language(body): |
|
55
|
|
|
"""return scores for detected languages formatted according to Swagger spec""" |
|
56
|
|
|
|
|
57
|
|
|
scores = lang_detector(body.get("text"), tuple(body.get("candidates"))) |
|
58
|
|
|
return { |
|
59
|
|
|
"results": [ |
|
60
|
|
|
{"language": s[0] if s[0] != "unk" else None, "score": s[1]} |
|
61
|
|
|
for s in scores |
|
62
|
|
|
] |
|
63
|
|
|
} |
|
64
|
|
|
|
|
65
|
|
|
|
|
66
|
|
|
def _suggestion_to_dict(suggestion, subject_index, language): |
|
67
|
|
|
subject = subject_index[suggestion.subject_id] |
|
68
|
|
|
return { |
|
69
|
|
|
"uri": subject.uri, |
|
70
|
|
|
"label": subject.labels[language], |
|
71
|
|
|
"notation": subject.notation, |
|
72
|
|
|
"score": suggestion.score, |
|
73
|
|
|
} |
|
74
|
|
|
|
|
75
|
|
|
|
|
76
|
|
|
def suggest(project_id, body): |
|
77
|
|
|
"""suggest subjects for the given text and return a dict with results |
|
78
|
|
|
formatted according to Swagger spec""" |
|
79
|
|
|
|
|
80
|
|
|
try: |
|
81
|
|
|
project = annif.registry.get_project(project_id, min_access=Access.hidden) |
|
82
|
|
|
except ValueError: |
|
83
|
|
|
return project_not_found_error(project_id) |
|
84
|
|
|
|
|
85
|
|
|
try: |
|
86
|
|
|
lang = body.get("language") or project.vocab_lang |
|
87
|
|
|
except AnnifException as err: |
|
88
|
|
|
return server_error(err) |
|
89
|
|
|
|
|
90
|
|
|
if lang not in project.vocab.languages: |
|
91
|
|
|
return connexion.problem( |
|
92
|
|
|
status=400, |
|
93
|
|
|
title="Bad Request", |
|
94
|
|
|
detail=f'language "{lang}" not supported by vocabulary', |
|
95
|
|
|
) |
|
96
|
|
|
|
|
97
|
|
|
limit = body.get("limit", 10) |
|
98
|
|
|
threshold = body.get("threshold", 0.0) |
|
99
|
|
|
|
|
100
|
|
|
try: |
|
101
|
|
|
hit_filter = SuggestionFilter(project.subjects, limit, threshold) |
|
102
|
|
|
result = project.suggest(body["text"]) |
|
103
|
|
|
except AnnifException as err: |
|
104
|
|
|
return server_error(err) |
|
105
|
|
|
|
|
106
|
|
|
hits = hit_filter(result).as_list() |
|
107
|
|
|
return { |
|
108
|
|
|
"results": [_suggestion_to_dict(hit, project.subjects, lang) for hit in hits] |
|
109
|
|
|
} |
|
110
|
|
|
|
|
111
|
|
|
|
|
112
|
|
|
def _documents_to_corpus(documents, subject_index): |
|
113
|
|
|
corpus = [ |
|
114
|
|
|
Document( |
|
115
|
|
|
text=d["text"], |
|
116
|
|
|
subject_set=SubjectSet( |
|
117
|
|
|
[subject_index.by_uri(subj["uri"]) for subj in d["subjects"]] |
|
118
|
|
|
), |
|
119
|
|
|
) |
|
120
|
|
|
for d in documents |
|
121
|
|
|
if "text" in d and "subjects" in d |
|
122
|
|
|
] |
|
123
|
|
|
return DocumentList(corpus) |
|
124
|
|
|
|
|
125
|
|
|
|
|
126
|
|
|
def learn(project_id, body): |
|
127
|
|
|
"""learn from documents and return an empty 204 response if succesful""" |
|
128
|
|
|
|
|
129
|
|
|
try: |
|
130
|
|
|
project = annif.registry.get_project(project_id, min_access=Access.hidden) |
|
131
|
|
|
except ValueError: |
|
132
|
|
|
return project_not_found_error(project_id) |
|
133
|
|
|
|
|
134
|
|
|
try: |
|
135
|
|
|
corpus = _documents_to_corpus(body, project.subjects) |
|
136
|
|
|
project.learn(corpus) |
|
137
|
|
|
except AnnifException as err: |
|
138
|
|
|
return server_error(err) |
|
139
|
|
|
|
|
140
|
|
|
return None, 204 |
|
141
|
|
|
|
NatLibFi /
Annif
